18 August 2020

Artificial Intelligence and Democratic Norms


The Sharp Power and Democratic Resilience series aims to contextualize the nature of sharp power, inventory key authoritarian efforts and domains, and illuminate ideas for non-governmental action that are essential to strengthening democratic resilience.

Dr. Nicholas D. Wright is an affiliated scholar at Georgetown University, an honorary senior research fellow at University College London, a consultant at Intelligent Biology, and fellow at New America. His work combines neuroscientific, behavioral, and technological insights to understand decision making in politics and international confrontations in ways practically applicable to policy.

This report discusses how to establish democratically accountable rules and norms that harness the benefits of artificial intelligence-related technologies, without infringing on fundamental rights and creating technological affordances that could facilitate authoritarian concentration of power. Absent these purposeful efforts, societies risk spiraling into new authoritarian forms of surveillance-based governance. Civil society around the world has a crucial role to play in helping democracies resist authoritarian pressure on the global surveillance environment. Organizations focused on diverse issues including privacy, human rights, free expression, technological standards, public health, and consumer protection can help identify, explain, and collaboratively address the complex challenges that arise from AI-related technologies.


Digital authoritarian competitors stand ready to exploit a lack of foresight in democracies and manipulate the development of global surveillance to serve their own interests.

KEY IDEAS

Building and maintaining data silos. Authoritarian regimes can turbocharge AI by training it on two types of data that liberal democracies should not similarly exploit or combine: “broad data” generated at volume on digital devices, and high quality “ground truth data,” such as tax returns and medical records. While conventional wisdom says that data must be integrated rather than isolated, siloing data limits authoritarian affordances and enhances security. Civil society must consider what silos are necessary to prevent misuse of data.

Affording new models of “digital sovereignty” for use by liberal democracies. Authoritarian states advocate for digital sovereignty as a state-based model of control over the internet. There is a critical need to develop alternatives. Civil society can help think through new models that balance sovereignty with the protection of individual freedoms.

Support tech–civil society collaborations and develop resilience. Civil society, in cooperation with government and big tech corporations where possible, can aim to correct market failures—like privileging advertising and marketing tools over individual privacy—by giving citizens the means to safeguard democratic integrity against malign information operations, while preserving essential openness of the information environment.

Resist sharp power in international fora. Norm-setting and technical standardization of AI-related technologies happen at a global scale. Civil society should promote transparent, multistakeholder AI governance and develop AI standards that encourage democratic practices and individual privacy.

No comments: